Recently I have been thinking about the generally accepted statistic that 80% of IT projects fail. I can't say this has been my experience. The more I though about it the more I came to the conclusion that this statistic is another example of the way the research houses like to sensationalise their research. A more dignified term is market. Startling "facts" like this of course justifies the high costs of their services. Why would you pay all that money if the report told you what you already know?

Is failure less than 50%?

So what is it that everyone already knows. We all know that 80% of IT projects don't meet 100% of their original goals and are not within budget. The research houses present IT projects success or failure as a binary outcome. How do they measure success or failure? Is success 50% of goals met and failure less than 50%? If so, I would say most IT projects are "passes", maybe not with distinction, but still not failures.

Most projects deliver value to the company

The reality is most projects deliver some value to the company that implemented them. The plethora of systems out there is testament to this. Of course the whys and the wherefores for failure to meet 100% of goals is well documented and I agree mostly with what the accepted industry wisdom is on this, although in my opinion there is lack of analysis on the role of modern corporate structure and practices in project failure.

You should tally your projects marks!

So if you want to know if your project is a success or failure you should agree up-front what the benefits and goals of the system are and, post implementation, mark off what has and what has not been achieved. You can then tell if you project is a child prodigy, average or  needs to go to special school. Don't just accept the fate of project failure prophisized by the research houses.